Empirical Null Distribution for Gamma Statistics with Application to Multiple Testing in RNA-Seq Experiments

• Activity Number: 340
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Xing Ren*+ and Jeffrey Miecznikowski and Jianmin Wang and Song Liu
• Companies: and University at Buffalo and Roswell Park Cancer Institute and Roswell Park Cancer Institute
• Keywords: empirical null distribution ; RNA-Seq ; Gamma mixture

Abstract

Genome and transcriptome platforms like microarrays and RNA-Seq often involve simultaneous hypothesis testing of thousands of genes or transcripts. A key step to control type I errors in such large-scale testing is to obtain the null distribution of the test statistics. We propose a Gamma mixture model for the test statistics of RNA-Seq data, and show by examples that the asymptotic null distribution is often inappropriate. Instead we propose new methods to estimate an empirical null distribution. The methods are evaluated via simulations and applied to real RNA-Seq dat sets.